適用於單載波區塊傳輸系統之應用訊號功率比之取樣時機及載波頻率追蹤演算法及改良的格雷碼相關器設計

Abstract

802.15.3c 是一個新的無線室內個人網路技術。其操作在60Ghz的免認證頻帶並提供至少1.5Gbps的資料率以滿足如高品質的影像傳輸的應用 在802.15.3c的2種傳輸模式: OFDM 以及SCBT 模式中，本篇論文選擇SCBT模式做為主要的研究主題，並在此前提之下，設計基頻接收機已符合規格要求以及其操作環境。 本篇論文主要內容集中在同步的議題，包括載波的同步以及取樣頻率的同步。首先本文提出一改良式的Golay匹配器，適用在15.3c所使用的π/2 BPSK 調變上，並利用此改良式的Golay匹配器，先實行粗略的載波同步。此改良式的Golay匹配器跟傳統的匹配器比起來，需要的運算量約為傳統的10分之一，可以大幅的降低複雜度。在估測殘餘的載波頻率以及取樣頻率的飄移上，為了使15.3c系統的資料率達到最大，本文不使用額外增加領航符元的方法，而是利用解碼後的資料與經過通道等化過後的資料，利用其相位差與取樣頻率以及載波頻率的關係，來估測載波頻率及取樣時機的飄移。其大概的程序為，先設一個能量的門檻來篩選一定數量且訊雜比較高的訊號，來減少雜訊的影響並使所需要的運算量減少，並根據訊號的能量給予不同的權重來更進一步的增加估測的準確度，估測取樣時機以及載波頻率的偏移。而在取樣時機偏移的補償上，本文利用模擬發現8倍的內插重建取樣訊號可達到很好的補償效能，並比較了不同的內插濾波器架構在補償上的效果，包括了1級架構的內插、先內插2倍再內插4倍的2級架構、以及每階內插2倍的3級架構，並由此選出較適用於15.3c 系統的架構。利用所提出的同步以及補償方法，經由模擬驗證，可得到接近理想狀況下的效能。

802.15.3c is a new indoor wireless communication technique. It operates at the 60Ghz unlicensed band and provides high data rates of more than 1.5 Gbps to satisfy the command of high-speed data transmission such as high definition video stream transmission. There are two transmission modes in 802.15.3c: OFDM mode and SCBT mode. The sampling timing/carrier frequency tracking part of SCBT mode has been designed in this thesis according to the frame format and the characteristics of the channel environment at 60GHz. The main contents in this thesis are carrier frequency offset (CFO) estimation, sampling timing offset estimation and compensation. For CFO estimation, an effective method based on Golay code is proposed to estimate frequency offset. It has low complexity compared to traditional correlation method. In sampling timing and carrier frequency offset tracking problem, unlike conventional tracking method in time domain, here we propose a frequency domain timing offset estimation methods. The proposed methods using the data after equalization and decision to detect their phase difference. By the phase difference, sampling timing and carrier frequency offset could be estimated. The offset estimation is started by using a threshold to select a number of signal samples in frequency domain which have higher SNR than most of received signal samples Second, we use the power of the selected signal samples as a weighting factor to enhance the estimation accuracy. The compensation is done by interpolating the new data sample base on the new sampling time instant. We found out that 8-time interpolation has very good performance by simulation. Here we compare different interpolation structure, including 1, 2 and 3stages design and two design methods: Least square and Equiripple. By adopting these interpolation designs to real tracking application, one can select suitable designs in 15.3c system. With the timing offset estimation method, high-accuracy timing offset can be estimated and compensated so as to maintain the system performance.